Molecular and Cellular Probes 22 (2008) 139–150

Contents lists available at ScienceDirect

Molecular and Cellular Probes

journal homepage: www.elsevier.com/locate/ymcpr

Distribution and quantification of Candidatus Liberibacter americanus,

agent of huanglongbing disease of citrus in São Paulo State, Brasil,
in leaves of an affected sweet orange tree as determined by PCR
Diva C. Teixeira a, b,1, Colette Saillard b, Carole Couture b, Elaine C. Martins a, Nelson A. Wulff a,
Sandrine Eveillard-Jagoueix b, Pedro T. Yamamoto a, Antonio J. Ayres a, Joseph M. Bové b, *
a
Fundecitrus, Av. Dr. Adhemar Pereira de Barros, 201, CEP 14807-040 Araraquara, SP, Brazil
b
Universite de Bordeaux 2, IBVM, UMR GDPP, Centre INRA, Institut National de la Recherche Agronomique, 71, Av. Edouard Bourlaux, 33883 Villenave d’Ornon Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history: Huanglongbing (HLB), an insect-transmitted disease of citrus, known for many years in Asia and Africa,
Received 30 July 2007 has appeared in the state of São Paulo State (SSP), Brazil, in 2004, and the state of Florida, USA, in 2005.
Accepted 14 December 2007 HLB endangers the very existence of citrus, as trees infected with the bacterial pathogen, irrevocably
Available online 26 February 2008
decline. In the absence of curative procedures, control of HLB is difficult and only based on prevention.
Even though not available in culture, the HLB bacterium could be shown to be Gram-negative and to
Keywords:
represent a new candidate genus, Candidatus Liberibacter, in the alpha subdivision of the Proteobacteria.
Citrus
Three Candidatus (Ca.) L. species occur: Ca. L. africanus in Africa, Ca. L. asiaticus in Asia, SSP, and Florida,
Huanglongbing
Liberibacter americanus and Ca. L. americanus in SSP. The liberibacters occur exclusively in the phloem sieve tubes. On affected
Blotchy mottle trees, HLB symptoms are often seen on certain branches only, suggesting an uneven distribution of the
Conventional PCR Liberibacter. Occurrence of Ca. L. americanus, the major HLB agent in SSP, has been examined in 822 leaf
Quantitative real time PCR samples from an affected sweet orange tree by two conventional PCR techniques and a newly developed
Brazil real time (RTi) PCR, also used for quantification of the Liberibacter in the leaves. Even though RTi-PCR
was able to detect as few as 10 liberibacters per gram of leaf tissue (l/g), no liberibacters could be de-
tected in any of the many leaf samples from a symptomless branch, while in blotchy mottle leaves
from symptomatic branches of the same tree, the Liberibacter titer reached values as high as 107 l/g.
These results demonstrate the uneven distribution of the Liberibacter in HLB-affected trees.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction
Huanglongbing (HLB) or yellow-shoot-disease is a destructive,
newly emerging, century-old malady of citrus [1]. Already
mentioned in southern China in the late 19th century, the disease
has been present for many years in three major regions of the
world: (i) Asia, including the Indian subcontinent, China, and South
East Asia, (ii) East Africa and Cameroun in West Africa, as well as
Madagascar, Reunion and Mauritius islands, and (iii) the south-
western Arabian Peninsula. HLB has now been reported for the first
time in two of the major citrus growing regions of America: São
* Corresponding author. Tel.: þ33 670 774 883; fax: þ33 557 122 369.
E-mail addresses: diva@fundecitrus.com.br (D.C. Teixeira), saillard@bordeaux.
inra.fr (C. Saillard), ccouture@bordeaux.inra.fr (C. Couture), laboratorio@fundecitrus.
com.br (E.C. Martins), nelsonwulff@fundecitrus.com.br (N.A. Wulff), jagoueix@
bordeaux.inra.fr (S. Eveillard-Jagoueix), ptyamamoto@fundecitrus.com.br (P.T. Yama-
moto), ayres@fundecitrus.com.br (A.J. Ayres), joseph.bove@wanadoo.fr (J.M. Bové).
1
Tel.: þ55 16 33 01 70 33; fax: þ55 16 33 01 70 41.
Paulo State, Brazil, and Florida State, USA, respectively, in 2004
and 2005. Today, only the Mediterranean basin, the Middle East,
Australia and New Zealand, and North- and South-Pacific islands
are still free of the affection. HLB is feared worldwide, because
citrus trees, once infected, will irrevocably decline. There are no
effective curative procedures, and thus control of HLB consists of
preventing trees from becoming infected.
Many features of HLB were already known when the disease
entered America. Lin Kung Hsiang of the South China agricultural
University transmitted the disease by graft-inoculation, and thus
proved for the first time the infectious nature of the disease agent
[2]. However, his results remained essentially unknown to the
western World for almost 10 years. In 1965, McClean and Ober-
holzer [3,4] confirmed the work of Lin by transmitting South
African HLB (greening) by graft-inoculation, and they showed the
African citrus psyllid, Trioza erytreae, to be a vector of the HLB agent
in South Africa. The Asian citrus psyllid, D. citri, was found to be
a vector of the HLB agent in 1967 in India [5] as well as in the
Philippines [6]. Biological control of the two psyllid vectors by

0890-8508/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mcp.2007.12.006
140 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150
Tamarixia spp. was achieved in Reunion Island [7,8]. The most
characteristic leaf symptom of HLB, blotchy mottle, well known in
China [2], was also stressed in other parts of the world [9–11].
The HLB agent, seen by electron microscopy for the first time in
1970 by Laflèche and Bové [12], is a Gram-negative bacterium
[13,14]. The HLB bacterium is exclusively restricted to the phloem
sieve tubes, not only in citrus, but also in two non-citrus,
experimental hosts, periwinkle (Catharanthus roseus) and tobacco
(Nicotiana xanthi) plants, to which the HLB bacterium was transmit-
ted from HLB-affected citrus by dodder (Cuscuta campestris) [10,15].
The HLB agent has never been obtained in culture, and had to be
characterized by DNA-based techniques. Comparisons of 16S
ribosomal DNA (16S rDNA) sequences [16], as well as 16S/23S inter-
genic region sequences [17], confirmed that the HLB agent was
a Gram-negative bacterium and, more precisely, a member of
a new lineage in the alpha subdivision of the Proteobacteria. The
trivial name liberobacter, later replaced by Liberibacter [18], was
given to the HLB bacterium. The Liberibacter strains in Africa could
be distinguished from those in Asia on the basis of temperature
sensitivity, DNA hybridizations and genomic properties, and
serology. For these reasons, they represent two different species:
Candidatus L. africanus in Africa and Candidatus L. asiaticus in
Asia, the term Candidatus (Ca) indicating that the organism has
not been cultured and has been characterized essentially by DNA-
based techniques [19]. A subspecies of the African liberibacter, Ca.
L. africanus subsp. capensis, has been characterized in an ornamen-
tal rutaceous tree, Calodendron capense, in the Western Cape prov-
ince of South Africa, where Ca. L. africanus was detected in 1996
[18,20]. In São Paulo state, a third liberibacter species, Ca. L. amer-
icanus, has been identified in 2004 [21–23]. The nusG–rplKAJL–
rpoBC gene clusters of the Asian and African liberibacters were
obtained and sequenced [24,25]. The same gene cluster has re-
cently been obtained from the American liberibacter (Teixeira, un-
published). Additional liberibacter genes, including the omp gene,
have been isolated by using Random Amplified Polymorphic DNA
(RAPD) [26]. The omp gene was used to study the genetic variability
of the Asian liberibacter [27]. Several monoclonal antibodies (MAs)
recognizing Asian or African strains of the HLB agent have been
obtained [28,29]. They revealed the existence of various serotypes
of the HLB agent, but they were too strain specific for detection
purposes [11,28]. MA 10A6, against an epitope of an Asian HLB
strain (Poona strain, India), was used in immunoaffinity chroma-
tography to purify the Poona HLB bacterium and visualize it for
the first time out of its cellular habitat. The length of the filamen-
tous bacteria varied from 1 to 4 m, with an average length of 2 m
and an average width of 0.2 m [28,30].
From 1970 to the early 1990s, electron microscopy (EM) was the
first, reliable technique to detect the HLB bacterium in suspicious
citrus trees and to confirm HLB symptoms [11,31]. The first molec-
ular detection technique was DNA hybridization. DNA probe In-2.6
[32] was specific for the Asian liberibacter, and DNA probe As-1.7,
for the African liberibacter [25]. EM and DNA hybridization have
been used to detect the liberibacters in several hundred leaf
samples from about 30 different countries [11,31,33,34]. Only the
African liberibacter was detected in Africa, and only the Asian
liberibacter in Asia. Liberibacters were seen by EM exclusively in
the sieve tubes. Probe In-2.6 was also used to detect the Asian
liberibacter on crush blots of Diaphorina citri psyllids [33].
The first PCR method for detection of the HLB liberibacters
became available in 1996, and was based on 16S rDNA amplification
with forward primers OI1 and OA1, and reverse primer OI2c [35].
With these primers, no amplification was obtained with the novel
American liberibacter, and new primers, f-GB1 and r-GB3, specific
for 16S rDNA amplification of the American HLB agent, were
developed [22]. A second PCR detection method, based on the
sequence of the rplKAJL gene cluster coding for ribosomal proteins
K, A, J, and L, was worked out for the detection of the African and
the Asian liberibacters [36]. The two PCR methods gave similar
results, in particular with 95% of 428 samples tested from Indone-
sia, South Africa and Mauritius Island.
HLB has been reported in the state of São Paulo (SSP) in March
2004. The disease in Brazil was found to be caused by two liberi-
bacters: Ca. L. asiaticus [22,37], present in about 4% of the trees,
and a new liberibacter, Ca. L. americanus [22], which infected
most of the trees; some trees were infected simultaneously with
the two liberibacters. Since 2006, an increase in the percentage of
trees infected with the Asian liberibacter has been witnessed. An
HLB control program has been developed in SSP, and conventional,
16S rDNA-based PCR has been used to confirm that trees with HLB
symptoms were indeed infected with the HLB agent(s). The PCR test
has been applied to several thousand samples of symptomatic
leaves from many different citrus farms. Practically, all leaves
with blotchy mottle symptoms gave positive PCR reactions. How-
ever, many HLB-affected trees do not show symptoms uniformly;
some branches having symptomatic leaves and fruits, others being
free of symptoms. Also, leaves with blotchy mottle symptoms are
not always available, and the possibility of using leaves with other
symptoms, such as zinc deficiency patterns, or even symptomless
leaves, should be envisaged. Finally, detection of the liberibacters
in recently infected trees, but showing no symptoms yet could be
hampered by irregular distribution of the HLB agent. Therefore,
we have studied the distribution of Ca. L. americanus, the major lib-
eribacter in SSP, in the various branches of a ‘‘Westin’’ sweet orange
tree by collecting 822 leaf samples, and analysing the samples for
liberibacter detection by conventional PCR, nested PCR, and a newly
developed SYBR Green real time PCR (RTi-PCR) with quantification
for Ca. L. americanus. The results of this work are presented here.
2. Materials and methods
2.1. Plant material, insects, and DNA extraction
Westin sweet orange leaves with blotchy mottle symptoms
were collected on several trees from an orchard in the Araraquara
region of the state of São Paulo (SSP), and used to identify the
infecting liberibacter species by conventional 16S rDNA PCR with
primer pair GB1/GB3 for Ca. L. americanus and primer pair OI1/
OI2c for Ca. L. asiaticus [22].
One 4-year-old, HLB-affected, Westin sweet orange tree infected
with Ca. L. americanus was selected for studying the distribution
and quantification of the liberibacters in the leaves throughout
the canopy. The position of each leaf sample was recorded and
materialized on the tree with a label carrying the sample number
(Fig. 1). A total of 822 leaf samples were collected. For each sample,
only leaves with one type of symptom were taken, and the type of
leaf symptom (blotchy mottle, zinc deficiency, no symptoms) was
recorded.
Leaves from a given sample (w5–10 leaves) were washed under
running tap water, blotted dry on filter paper, and their midribs
were recovered and cut into small pieces. Five hundred milligram
of midribs were ground in a Homex 6 homogenizer (Bioreba AG,
CH-4153 Reinach BL1, Switzerland). Total DNA was extracted using
the CTAB (cetyl trimethyl ammonium bromide) method [38]. DNA
preparations were adjusted at 100 ng DNA/ml and kept at 20  C.
Healthy leaf DNA was prepared in Bordeaux (France) from
greenhouse-grown, sweet orange seedlings.
Adults of Diaphorina citri, the psyllid insect vector of HLB in SSP,
were collected on trees with or without symptoms of HLB, and used
for DNA extraction by the CTAB method. The DNA preparations
from 105 batches of five psyllids each were adjusted with water
to a final volume of 30 ml and used in conventional rpl-PCR with
new rpl primers (see 2.3.1) and in RTi-PCR.
 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150
Fig. 1. HLB-affected, four-year-old Westin sweet orange tree infected with Ca.
L. americanus. The position of each one of the 822 leaf samples collected on the tree
was marked by a white label. The central zone of the photo shows the most affected
part of the tree.
2.2. Primers for detection of Ca. L. americanus
and/or Ca. L. asiaticus by conventional PCR
2.2.1. 16S rDNA PCR for Ca. L. americanus or Ca. L. asiaticus
The two primer pairs GB1/GB3 and OI1/OI2c for 16S rDNA
amplification were used for routine PCR detection of Ca. L. ameri-
canus and Ca. L. asiaticus, respectively, according to Teixeira et al.
[22], except that for Ca. L. americanus detection, the annealing
temperature was decreased from 64 to 62  C.
2.2.2. Nested 16S rDNA PCR for Ca. L. americanus
16S rDNA of Ca. L. americanus was amplified with universal
primers FD1 and RP1 for bacterial 16S rDNA [39]. The reaction
mixture contained in a final volume of 40 ml: 1PCR buffer (Prom-
ega Corporation, Madison, WI, USA), 2 mM MgCl2, 0.2 mM of each
dNTP, 500 nM of each primer (FD1 and RP1), 1.5 U of Taq DNA
polymerase (Promega), and 5 ml of DNA preparation. The program
consisted of an initial step at 94  C for 3 min, and 35 cycles of
94  C for 1 min, 52  C for 1 min, 72  C for 90 s, and a final extension
at 72  C for 7 min. The PCR product was diluted 10 times with water,
and 1 ml of the dilution was used for 16S rDNA amplification with
Ca. L. americanus-specific primers GB1 and GB3 [22]. The Westin
sweet orange leaf samples negative for Ca. L. americanus by con-
ventional 16S rDNA PCR were analyzed by the nested PCR protocol.
2.3. Conventional PCR and RTi-PCR detection of Ca.
L. americanus and Ca. L. asiaticus with new rpl primers
2.3.1. Designation of new rpl primers
The sequence of the rplKAJL-rpoBC gene cluster (b operon) of Ca.
L. asiaticus is available since 1993 [24] (GeneBank accession
number: M94319). The corresponding sequence of Ca. L. ameri-
canus has only been determined recently (Teixeira, unpublished).
The two sequences were used to design new, specific primers for
PCR detection of the two respective liberibacters, using the Beacon
Designer Version 5 software (Premier Biosoft International).
Primers for Ca. L. asiaticus were designed from the rplL gene; the
sequences of forward primer f-rplLAs and reverse primer r-rplLAs
were 5’CGCCCGTTTCCGTTGT 3’ and 5’AGCCTCTTTAAGCCCTAAAT-
CAG 3’, respectively, and led to an amplicon of 161 bp. Primers for
Ca. L. americanus were from the rplJ gene; forward primer f-rplJAm
and reverse primer r-rplJAm, giving an amplicon of 127 bp, had,
respectively, the following sequences:
5’GGACAAGGGGATATTGGATAATGATG3’ and 5’ATTAAGAGTTCTA
AGCAACCTGACAG3’.
141
The two primer pairs were used in conventional rpl-PCR and in
SYBR Green RTi-PCR.
2.3.2. Conventional rpl-PCR with new rpl primers
Conventional rpl-PCR was carried out using an iCycler PCR DNA
System (Biorad). The best annealing temperatures and primer
concentrations were determined on DNA preparations from Ca. L.
asiaticus- or Ca. L. americanus-infected leaves. DNA preparations
from healthy sweet orange leaves and water were used as negative
controls. These assays resulted in the following conditions. With
each primer pair, 40 cycles were performed preceded by a 3 min
denaturation step at 94  C and followed by an elongation step for
5 min at 72  C. Cycle conditions were as follows: 30 s at 95  C,
30 s at 67  C and 30 s at 72  C. Reaction mixture of 25 ml contained
500 ng total DNA, 150 mM of each rpl primers, 200 mM dNTP,
300 mM MgCl2 and 1.25 U of Taq DNA polymerase (Promega) in
the reaction buffer supplied by the manufacturer.
2.3.3. SYBR Green RTi-PCR
For RTi-PCR, the AbsoluteÔ QPCR SYBRÒ Green Fluorescein Mix
(master mix, ABgene) was used. The SYBR Green reaction was
performed in a 25-ml reaction mixture containing 1  master mix,
0.15 mM of each primer, and 500 ng of total DNA preparation. The
iCycler IQ System (BioRAD) was used with the following program
for DNA amplification: 95  C for 15 min, 40 cycles each at 95  C
for 30 s, 67  C for 30 s, 72  C for 30 s, and a final extension at
72  C for 10 min. For detection of Ca. L. americanus, the SYBR Green
RTi-PCR was carried out with primer pair f-rplJAm/r-rplJAm, and
with primer pair f-rplLAs/r-rplLAs for Ca. L. asiaticus. After the
PCR reaction, the temperature was increased by 0.5  C every 10 s
from 69 to 95  C with a continuous measurement of fluorescence,
to obtain the melting curve data. Each DNA sample tested in
RTi-PCR was in duplicate or triplicate. The amplification curves
for Ct (threshold cycle) determination, and melting curves for
temperature estimations at the peak of the curves, were analyzed
using the iCycler software version 3.0.6070 from BioRad.
2.3.4. Specificity of SYBR Green RTi-PCR
The specificity of the new rpl primer pairs f-rplJAm/r-rplJAm and
f-rplLAs/r-rplLAs in SYBR Green RTi-PCR was evaluated with DNA
preparations from axenic cultures of Xylella fastidiosa, the agent of
citrus variegated chlorosis, Xanthomonas axonopodis pv. citri, the
agent of citrus canker, and Spiroplasma citri, the agent of citrus stub-
born disease, as well as with DNA preparations from lime (Citrus
aurantifolia) leaves infected with Candidatus Phytoplasma auranti-
folia, the agent of witches’ broom disease of lime. DNA preparations
from healthy sweet orange leaves and symptomatic sweet orange
leaves infected with Ca. L. asiaticus or Ca. L. americanus were
used as controls in addition to water. Samples were run in triplicate.
2.3.5. Quantitative RTi-PCR for Ca. L. americanus
The pGEM-T Easy vector containing a 1704 bp DNA fragment
amplified from the rplKAJL – rpoB gene cluster of Ca. L. americanus
(plasmid pDCT) was used to construct the calibration curve for
quantitative RTi-PCR. The 1704 bp insert contained the 127 bp
target DNA sequence amplified with primer pair f-rplJAm/r-rplJAm.
Eight micrograms of recombinant plasmid were linearized with
restriction enzyme Pst1 at 37  C during the night. The DNA was
precipitated with 2 volumes of 95% ethanol and 0.1 volume of
3 M NaOAc, washed with 70% ethanol, and dissolved in water.
The DNA concentration was determined spectrophotometrically,
and the DNA concentration was adjusted to 100 ng/ml.
Knowing the number of plasmid molecules in 5 ml, a 1:10
dilution series has been made to construct the calibration curve
and calculate the PCR reaction efficiencies with plasmid DNA for
quantitative RTi-PCR.
142 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150

We have used the term Limit of Quantification (LOQ), for assay
sensitivity, and the term Limit of Detection (LOD) for sample limit
of detection [40]. In our case, the LOQ (gene copies per reaction)
was calculated from the calibration curve to determine the lowest
concentration of bacterial genomes that remained within the linear
range of quantification [41].
3. Results
3.1. Evaluation of conventional rpl-PCR and SYBR Green
RTi-PCR for detection of Ca. L. americanus and
Ca. L. asiaticus. Quantification of Ca. L. americanus
3.1.1. Conventional rpl-PCR with new rpl primers
PCR products with DNA from sweet orange leaves infected with
Ca. L. americanus or Ca. L. asiaticus had the expected sizes of 127 bp
and 161 bp, respectively. No amplification was obtained with water,
DNA from healthy sweet orange leaves, or when primers specific for
Ca. L. americanus were used with Ca. L. asiaticus DNA, and vice-
versa (data not shown). Also, with either one of the two rpl primer
pairs, amplification was negative with DNA preparations from
sweet orange leaves infected with Xylella fastidiosa, or lime leaves
infected with Candidatus Phytoplasma aurantifolia. Similarly, the
PCR reaction was negative with DNA from an axenic culture of X.
axonopodis pv. citri, or Spiroplasma citri (data not shown).
3.1.2. SYBR Green RTi-PCR
With blotchy mottle leaf sample Am4 infected with Ca. L.
americanus alone, and using the rpl primers specific for Ca. L. amer-
icanus (f-rplJAm and r-rplJAm), the two PCR amplification curves
for duplicated assays were overlapping and had a mean Ct value
of 24.89 (Fig. 2, A1, blue curves). Similarly, with leaf sample As6
infected with Ca. L. asiaticus alone, and using the rpl primers
specific for Ca. L. asiaticus (f-rplLAs and r-rplLAs), the two PCR
amplification curves had a mean Ct value of 21.1 (Fig. 2, B1, brown
curves). Fig. 2, A2 and B2, shows the corresponding melting curves
with peaks at 82.0  C and 84.0  C for Ca. L. americanus and Ca. L.
asiaticus, respectively.
With blotchy mottle samples 2 and 121, each one being infected
with both liberibacters, amplification was obtained only with the
liberibacter species corresponding to the primer pair used in the
reaction mixture. With the primers for Ca. L. americanus, amplifica-
tion curves for samples 2 and 121 had mean Ct values of 26.55
(Fig. 2, A1, green curve) and 23.88 (Fig. 2, A1, red curve), respec-
tively, and the peak value of the melting curves (Fig. 2, A2) was
82  C, the value specific of Ca. L. americanus. With the primers for
Ca. L. asiaticus, the amplification curves for samples 2 (Fig. 2, B1,
pink curve) and 121 (Fig. 2, B1, black curve) had mean Ct values
of 23.5 and 24.3, respectively, and all melting curves (Fig. 2, B2)
had the same peak value, 84  C, the value specific for Ca. L. asiaticus.
In the absence of DNA (water control) or with DNA from healthy
sweet orange leaves, no Ct values were obtained. With samples
containing DNA from bacterial citrus agents other than the Ameri-
can and Asian liberibacters, namely Xylella fastidiosa, X. axonopodis
pv. citri, Spiroplasma citri, Candidatus Phytoplasma aurantifolia, and
Ca. L. africanus, melting curves non-specific of Ca. L. asiaticus and
Ca. L. americanus were obtained (data not shown).
3.1.3. Quantification of Ca. L. americanus in plants
Tenfold serial dilutions of plasmid pDCT from 500 fg/ml (105
copies/ml) to 0.005 fg/ml (1 copy/ml) were used. Amplification was
carried out with three repetitions for each dilution (Fig. 3). It

Fig. 2. Detection of Ca. L. americanus or Ca. L. asiaticus by SYBR Green RTi-PCR in samples infected by one or two liberibacters. A1 and A2: respectively, RTi-PCR amplification curves
and melting curves, with primers f-rplJAm/r-rplJAm, for samples ‘‘Am4’’ (infected with Ca. L. americanus alone, Ct mean value ¼ 24.89, blue line), ‘‘2’’ (infected with both
liberibacters, Ct mean value ¼ 26.55, green line), and ‘‘Am121’’ (infected with both liberibacters, Ct mean value ¼ 23.88, red line). Melting curve peak is at 82  C. B1 and B2: respec-
tively, RTi-PCR amplification curves and melting curves, with primers f-rplLAs/r-rplLAs, for samples ‘‘As6’’ (infected with Ca. L. asiaticus alone, Ct mean value ¼ 21.1, brown line), ‘‘2’’
(infected by both liberibacters, Ct mean value ¼ 23.5, pink line), and ‘‘121’’ (infected with both liberibacters, Ct mean value ¼ 24.3, black line). Melting curve peak is at 84  C. The
samples were run in duplicate.
 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150 143

Table 1
Quantification of Ca. L. americanus in different samples by the calibration curve
method

Sample Symptoms No. of bacteria in No. of bacteria

(designation) 500 ng of DNA (B) per gram of
midribs (N )
1 (Am4) Blotchy mottle 52,500 1.1  108
2 (121) Blotchy mottle 99,100 2.1  107
3 (2) Blotchy mottle 18,500 1.5  107
4 (86) Blotchy mottle 43,500 1.1  107
5 (90) Zn deficiency 30,400 8.7  106
6 (98) Zn deficiency 1580 4.3  105
7 (102) Blotchy mottle 1600 4.6  105
8 (107) Blotchy mottle 10,200 4.2  106
9 (118) Zn deficiency 10,600 2.3  106
10 periwinkle 1 Mild mottle 4,370 1.4  106
11 Periwinkle 2 Mild mottle 10,200 3.7  106
12 Periwinkle 3 Severe mottle 447,000 3.4  108
Fig. 3. RTi-PCR amplification curves for tenfold serial dilutions of plasmid pDCT,
ranging from 500 fg/ml (105 copies/ml) to 0.005 fg/ml (1 copy/ml). DNA preparations from leaf samples of nine different Ca. L. americanus-infected
sweet orange trees (1–9) in SSP, as well as from periwinkle leaves with mild
(10, 11) and severe symptoms (12) from the Bordeaux laboratory greenhouse,
made no difference whether dilutions were made with water or were used to estimate the number of liberibacters per g of fresh leaf midribs
(N ). The following formula was used to calculate N: N ¼ 2B  Q  2, with
a solution of DNA from healthy citrus midribs. The calibration curve
B ¼ number of bacteria in 500 ng of infected plant DNA, given automatically by
had an average slope value of 3.777, indicating that the efficiency the calibration curve method, and Q ¼ total quantity of DNA in mg extracted
of the PCR reaction was 84% (Fig. 4). The limit of detection (LOD) from 500 mg of fresh midribs.
and the limit of quantification (LOQ) were equivalent to one copy,
however, only in two of the three repetitions. Also, because the
efficiency of PCR amplification is only 84%, the most likely detect- ones (sample 79) was positive by RTi-PCR, but only when not
able limit is w10 liberibacters. The numbers of liberibacters in diluted. The number of liberibacters in this sample was low and
samples A and B, as deduced from the calibration curve, were, equivalent to about 100 bacteria per insect (Table 2). The eight
respectively, 1.61 102 and 2.47  105 liberibacters in 500 ng of samples found positive by rpl-PCR were also positive by RTi-PCR,
DNA preparation (Fig. 4). Several samples of symptomatic leaves even after dilution. The number of liberibacters in these insects
from different sweet orange trees in São Paulo state (Table 1, was much higher and varied from 1.0  103 to 6.8  104.
samples 1–9) as well as symptomatic periwinkle (C. roseus) leaves
from the Bordeaux laboratory greenhouse (Table 1, samples 3.1.5. Sensitivities of various PCR methods for the
10–12) were analyzed by the quantitative RTi-PCR. The number of detection of Ca. L. americanus
liberibacters in 1 g of fresh midribs (l/g) from blotchy mottle sweet Fig. 5 shows that all PCR methods used were able to clearly
orange leaves varied from 4.6  105 to 1.1 108 l/g, with an average detect 104 bacteria per reaction. With RTi-PCR, the last dilution,
value of 2.7  107 l/g; for leaves with zinc deficiency symptoms, the corresponding theoretically to one liberibacter (sample 7), gave
average value was 3.8  106 l/g. In periwinkle plants with mild a positive reaction with only two out of the three triplicates,
symptoms, the average number of liberibacters reached 2.5  suggesting that the detection limit was reached. Consequently
106 l/g of fresh midribs, while in midribs from leaves with severe with field samples, the sensitivity of the RTi-PCR is not much higher
symptoms, the value was 100 times higher. than that obtained with nested PCR, allowing the detection of 10
liberibacters in the reaction mixture. The sensitivities provided by
3.1.4. Quantification of Ca. L. americanus in psyllids RTi-PCR and nested PCR were 100 times higher than that of rpl-
The 105 psyllid DNA samples were first analyzed by conven- PCR and 1000 times higher than that of conventional 16S rDNA PCR.
tional rpl-PCR with primers f-rplJAm and r-rplJAm for Ca. L.
americanus, and primers f-rplLAs and r-rplLAS for Ca. L. asiaticus. 3.2. Distribution of Ca. L. americanus in leaf
Eight samples only gave positive amplifications, and only for Ca. samples from the Westin sweet orange tree
L. americanus (data not shown). Among the remaining 97 negative
samples, 15 were chosen randomly and were submitted with the 3.2.1. Results of PCR tests
eight positive samples to quantitative RTi-PCR before and after 10 A total number of 822 leaf samples (100%) were collected on the
fold dilution with water. A single sample among the negative Westin sweet orange tree. The samples with blotchy mottle leaves

Fig. 4. Quantitative RTi-PCR for Ca. L. americanus. Calibration curve from tenfold serial dilutions of pDCT plasmid, and quantification for samples A (1.6  102 liberibacters in 500 ng
of DNA) and B (2.5  105 liberibacters in 500 ng of DNA). The samples were in triplicate.
144 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150

Table 2
Quantification of Ca. L. americanus in Diaphorina citri psyllids
Sample
39
63
67
86
92
94
101
79
The number of bacteria in 500 ng of insect DNA, given automatically by the calibra-
tion curve method, was multiplied by the total quantity of DNA (in mg) extracted
from one insect.
numbered 111, those with zinc deficiency symptoms, 77, and those
without symptoms, 634 (Fig. 6). All 822 samples were first analyzed
by conventional 16S rDNA PCR (c-PCR) with primers GB1 and GB3.
The leaf samples giving positive c-PCR reactions amounted to 253
(31%), of which 111 (13.5%) showed blotchy mottle (Fig. 7A, samples
9, 12, and 14 for instance), 61 (7.4%) had zinc deficiency symptoms
(Fig. 7A, sample 6 for instance), and 81 (10%) were symptomless.
The 111 c-PCR-positive blotchy mottle leaf samples represented
the totality of the blotchy mottle samples.
The remaining 569 c-PCR-negative samples (Fig. 6) were
submitted to nested PCR (n-PCR). Out of the 34 samples giving
a positive n-PCR reaction (6.0%), 4 showed zinc deficiency symp-
toms and 30 were symptomless. Samples 81, 88, and 91 are such
symptomless leaf samples, negative by c-PCR (Fig. 7A), but positive
by n-PCR (Fig. 7B). Taking into account the results from both c-PCR
No. of bacteria and n-PCR, the positive samples reached 253 þ 34 ¼ 287 (34.9%) of
per insect which 111 (13.5%) presented blotchy mottle, 61 þ 4 ¼ 65 (7.9%) zinc
1.2  104 deficiency symptoms, and 81 þ 30 ¼ 111 (13.5%) no symptoms.
2.2  104 The samples that were negative by the above two PCR techniques
1.0  103
numbered 535, of which 12 had zinc deficiency symptoms, and 523
3.4  104
5.6  104 showed no symptoms (Fig. 6). Among these 535 samples, 97
7.4  103 samples were chosen, randomly with replacement, on the various
4.2  104 branches of the tree for RTi-PCR. Six of the 97 samples had zinc
9.6  101 deficiency symptoms, and 91 were symptomless. Only five samples,
all symptomless, gave positive amplifications with RTi-PCR. Samples
321, 335, 346, and 477 are four of the five symptomless samples,
c-PCR-negative, as well as n-PCR-negative (Fig. 7B), but RTi-PCR-
positive (Fig. 7C). If testing 97 randomly chosen samples gave 5
positive samples, testing all 535 samples could have given 28
putative samples, bringing the estimated number of PCR-positive,
symptomless leaf samples close to 81 þ 30 þ 28 ¼ 139 (17%).
Table 3 gives in detail the results from the above three PCR tech-
niques according to leaf symptoms (blotchy mottle, zinc deficiency
symptoms, without symptoms) for all branches and sub-branches.
The tree had essentially three major branches (1, 2, and 3) and two
minor sub-branches (4-Q and 5-A). Each major branch was com-
posed of several sub-branches 1-S to 1-AC, 2-B to 2-R, and 3-I to
3-P. The number of samples collected on each sub-branch (figure
in parentheses) and the sample designation (1–822) are also
indicated on Table 3. The position of each leaf sample on a given
sub-branch and the PCR results for each sample are illustrated for

A
Sample number 1 2 3 4 5 6 7

ng of infected DNA 1000ng 100ng 10ng 1ng 0,1ng 0,01ng 0,001ng

number of bacteria 1 x 106 1 x 105 1 x 104 1000 100 10 1

16S rDNA PCR (GB1/GB3) + + + +/- - - -

rpl PCR (f- rplJAm/r-rplJAm) + + + + +/- - -

16S rDNA nested PCR
+ + + + + + -
(FD1/RP1 - GB1/GB3)

RTi-PCR (f-rplJAm/r-rplJAm) + + + + + + +/-

16S rDNA PCR rpl PCR 16S rDNA nested PCR

(GB1/ GB3) (f-rplJAm/r-rplJAm) (FD1/RP1 - GB1/ GB3)

Rti-PCR
(f-rplJAm/r-rplJAm)

Fig. 5. Sensitivities of various PCR methods for detection of Ca. L. americanus. Four PCR methods have been compared: 16S rDNA PCR (primers GB1/GB3); rpl-PCR (primers f-
rplJAm/r-rplJAm); 16S rDNA nested PCR (primers FD1/RP1 – GB1/GB3); and RTi-PCR (primers f-rplJAm/r-rplJAm). DNA sample 12 (see Table 1) from midribs of periwinkle leaves
known to be infected with Ca. L. americanus was adjusted to an initial concentration of 200 ng of DNA/ml, and used to prepare a series of tenfold dilutions with water. Five ml of the
various DNA preparations were used for each PCR reaction. The number of liberibacters in 1000 ng of the initial DNA preparation was estimated by quantitative RTi-PCR to be
447,000  2 (Table 1), close to 106. The quantity of DNA per reaction ranged from 1000 ng or 106 bacteria to 0.001 ng or 1 bacterium. A: PCR results. þ, , þ/: positive, negative,
borderline PCR reaction, respectively. B: Agarose gel electrophoresis of amplicons from 16S rDNA PCR, rpl-PCR, and nested PCR, as well as RTi-PCR amplification curves. 1–7: sample
numbers; Cþ: positive control for Ca. L. americanus; H: negative control (DNA from healthy periwinkle leaves); O: negative control (water). M: 1 kb DNA ladder.
 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150 145

822 samples 569 samples 535 samples

111 with blotchy mottle

77 Zn deficiency
634 without symptoms

Tested by Tested by 97 samples / 535 (18%)

16S rDNA 16S nested
PCR PCR 91 no symptoms

06 Zn deficiency

253+ / 822 (31%) 34+ / 569 (6%)

111 with blotchy mottle 30 no symptoms Tested by

RTi-PCR
142 without blotchy mottle: 04 Zn deficiency

81 no symptoms
61 Zn deficiency 569 - 34 =535
05+ / 97 (4%)
569 - / 822 (69%) 535- :
523: no symptoms 05 no symptoms
without blotchy mottle
12: Zn deficiency

Fig. 6. Flow sheet for detection of Ca. L. americanus in leaf samples from Westin sweet orange tree by 16S rDNA PCR (c-PCR), 16S rDNA nested PCR (n-PCR), or RTi-PCR. All 822 leaf
samples from the tree were submitted to c-PCR. The 569 samples found negative by c-PCR were analyzed by n-PCR. After c-PCR and n-PCR, 535 samples remained negative, of which
97 were randomly chosen and submitted to RTi-PCR. Each sample had only one type of leaves: blotchy mottle leaves, zinc deficiency leaves, or leaves with no symptoms. PCR results
(number of positive samples) are indicated for each type of leaf sample. The number of samples analyzed, and the number of PCR-positive or PCR-negative samples are indicated.

Fig. 7. PCR detections of Ca. L. americanus in various types of leaf samples from the Westin sweet orange tree. Agarose gel electrophoresis of amplicons from c-PCR (A) on leaf
samples with blotchy mottle (samples 8, 12 and 14), Zn deficiency (sample 6), and on symptomless leaves (samples 81, 88 and 91), and n-PCR (B) on symptomless, c-PCR-negative
leaf samples 81, 88, 91, 321, 335, 346, 477. Amplification curves of RTi-PCR (C) with n-PCR-negative samples 321, 335, 346 and 477. O: water control. C-: negative control with DNA
from healthy sweet orange leaves. Cþ: positive control with DNA from Ca. L. americanus-infected sweet orange leaves. M ¼ 1 kb DNA ladder.
146 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150

Table 3
Detection of Ca. L. americanus by three PCR techniques in leaf samples from the various branches and sub-branches of Westin sweet orange tree

Branches, Samples per sub-branch Blotchy mottle Zn deficiency Without symptoms

sub-branches (number of samples)
c-PCR c-PCR n-PCR RTi-PCR c-PCR n-PCR RTi-PCR
S 607–614 (8) 0/8 0/8 0/3
1-T 615–626 (12) 0/12 0/12 0/2
1-U 627–666 (40) 0/40 0/40 0/3
1-V 667–673 (7) 0/7 0/7 0/3
1-W 674–706 (33) 0/33 0/33 0/3
1-X 707–750 (44) 0/44 0/44 0/3
1-Y 751–768 (18) 0/18 0/18 0/3
1-Z 769–781 (13) 0/13 0/13 0/3
1-AA 782–801 (20) 0/20 0/20 0/3
1-AB 802–811 (10) 0/10 0/10 0/3
1-AC 812–823 (12) 0/12 0/12 0/3
Total 217 0/217 0/217 0/32

2-B 38–95 (58) 4/4 15/17 0/2 0/1 7/37 4/30 1/5
2-C 96–139 (44) 2/2 5/6 0/1 6/36 0/30 0/5
2-D 140–176 186–195 (47) 8/8 1/2 0/1 8/37 0/29 0/6
2-E 177–185 (9) 0/1 0/1 0/1 0/8 1/8 0/2
2-F 196–212 (17) 0/1 0/1 0/1 0/16 0/16 0/4
2-G 213–225 (13) 0/13 0/13 0/5
2-H 226–237 (12) 6/6 4/4 1/2 0/1 0/1
2-R 593–606 (14) 7/7 5/5 0/2 0/2 0/1
Total 214 27/27 30/36 0/6 0/3 22/151 5/129 1/29

3-I 238 (1) 1/1

3-J 239–246 (8) 1/1 0/7 1/7 0/3
3-L 247–270 (24) 3/3 1/1 1/20 0/19 0/5
3-M 271–358 (87) 26/26 6/10 1/4 0/3 11/51 6/40 3/8
(- sample 299)
3-N 359–419 (61) 16/16 8/8 12/37 11/25 0/4
3-O 420–443 (24) 0/2 0/2 0/22 0/22
3-P 444–463 (20) 0/20 0/20
Total 225 47/47 15/21 1/6 0/3 24/157 18/133 3/20

4-Q 464–592 (129) 28/28 15/19 3/4 31/82 7/51 1/3

Total 129 28/28 15/19 3/4 31/82 7/51 1/3

5-A 1–37 (37) 9/9 1/1 4/27 0/23 0/7

Total 37 9/9 1/1 4/27 0/23 0/7

Total 822 111/111 61/77 4/16 0/6 81/634 30/553 5/91

The 822 samples included 111 samples with blotchy mottle leaves, 77 samples with zinc deficiency leaves, and 634 samples with leaves without symptoms. PCR techniques
used were: c-PCR (16S rDNA PCR with primer pair GB1/GB3), n-PCR (16S rDNA nested PCR with universal primer pair FD1/RP1 followed by primer pair GB1/GB3), and SYBR
Green RTi-PCR. Results are expressed as numbers of PCR-positive samples over total number of samples tested. For instance, 15/17 indicates that, on sub-branch 2-B (which has
58 samples numbered 38 to 95), there were 17 samples with zinc deficiency leaves, 15 of which were positive for Ca. L. americanus by c-PCR.

sub-branches 1-U and 1-V (Fig. 8A), 2-B (Fig. 8B), 3-M (Fig. 8C), 4-Q
(Fig. 8D), and 5-A (Fig. 8E). The presence of fruits, with or without
symptoms of HLB, is also indicated on the sub-branches of Fig. 8.
As a whole, these results show that none of the 217 samples
collected on symptomless branch 1 (Table 3) gave positive reactions
with any of the three PCR techniques. Regarding the symptomatic
branches, the cumulated number of samples positive by c-PCR,
n-PCR, and RTi-PCR amounted to 111 of 111 samples with blotchy
mottle leaves, and 65 of 77 samples with zinc deficiency symptoms.
In the case of symptomless leaves, the 634 samples came from the
whole tree, and the number of PCR-positive samples was estimated
at 139 (see above).
3.2.2. Quantification of HLB liberibacters in various leaf samples
The number of HLB liberibacters per gram of leaf midrib tissue
was determined on 43 samples from symptomatic branches
(Table 4): (i) 10 samples positive by c-PCR, including 6 samples
with blotchy mottle leaves, 2 samples with zinc deficiency leaves,
and 2 samples with symptomless leaves; (ii) 28 samples negative
by c-PCR, but positive by n-PCR, including 3 samples with zinc
deficiency leaves, and 25 samples with symptomless leaves; (iii) 5
samples with symptomless leaves, negative by c-PCR and n-PCR,
but positive by RTi-PCR. In addition, quantification was carried
out on 32 samples distributed randomly over all the sub-branches
of symptomless branch 1 (Table 3). All 32 samples from this branch
were found to be PCR-negative.
The highest concentrations of liberibacters were found in the
leaves with blotchy mottle, and ranged from 1.1 106 to 2.8  107
liberibacters per gram of tissue (l/g), averaging 8.6  106 l/g (Table
4, samples 1 (14) to 6 (167), and Fig. 7, samples 8, 12, 14). PCR-positive
leaves with zinc deficiency symptoms were of two types: (i) those
giving positive reactions with c-PCR, and they averaged 6.3  106 l/
g (Table 4, samples 7 (172) and 8 (6), and Fig. 7, sample 6), and (ii)
those negative by c-PCR, but positive by n-PCR, and they averaged
6  105 l/g [Table 4, samples 11 (323), 12 (468), and 13 (470)].
Some samples with symptomless leaves had also relatively high
liberibacter concentrations, with values around 106 l/g, averaging
2  106 l/g [Table 4, samples 9 (195), 10 (10), and 14 (363)].
However, many other samples with symptomless leaves had
much lower liberibacter concentrations. In the case of samples neg-
ative by c-PCR, but positive by n-PCR, they ranged from 2.2  103 l/g
[Table 4, sample 38 (378)] to 1.1 105 l/g [Table 4, sample 15 (361)],
averaging 2.4  104 l/g. The lowest liberibacter concentrations were
experienced with five symptomless samples, positive only by
RTi-PCR. [Table 4, samples 39 (56) to 43 (335)]. They averaged
4.6  102 l/g. The results of n-PCR and RTi-PCR of samples 321,
335, 346, and 477 [Table 4, samples 40 (346) to 43 (335)] are illus-
trated on Fig. 7, B and C.
 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150 147

Fig. 8. Five representative sub-branches of Westin sweet orange tree. A, B, C, D, and E: respectively, sub-branches 1-U/1-V, 2-B, 3-M, 4-Q, and 5-A. Red triangles: c-PCR-positive
samples with blotchy mottle leaves; red circles: c-PCR-positive samples with zinc deficiency leaves or symptomless leaves; blue circles: c-PCR-negative and n-PCR-positive samples
with zinc deficiency leaves or symptomless leaves; green circles: c-PCR-negative, n-PCR-negative, and RTi-PCR-positive samples; white circles: c-PCR-negative, n-PCR-negative
samples not tested by RTi-PCR; green rings: c-PCR-negative, n-PCR-negative, and RTi-PCR-negative samples; yellow ellipses: normal sweet oranges; green ellipses: HLB-affected
sweet oranges.

3.2.3. Distribution and quantification of liberibacters in and RTi-PCR, yields values of 106–107 l/g (samples indicated by
leaves of various sub-branches red triangles and circles on Fig. 8), 103–105 l/g (Fig. 8, blue circles)
On the basis of the above results, a conservative estimation of and 101–103 l/g (Fig. 8, green circles), respectively. On Fig. 8, leaf
the liberibacter titers in leaves testing positive by c-PCR, n-PCR samples testing negative by both c-PCR and n-PCR are indicated
148 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150

Table 4 and distal ends. Most symptomatic fruits were near PCR-positive
Quantitafication of Ca. L. americanus in various leaf samples from Westin sweet leaf samples. Finally, on sub-branch 2-B (Fig. 8B), there were
orange tree
practically no proximal samples, and the PCR-positive samples
Symptoms Sample Sub-branch c-PCR n-PCR RTi-PCR Bacteria were at the distal ends.
(designation) per gram
On the most affected parts of the sub-branches (Fig. 8, distal end
of midribs
of sub-branches 5-A and 4-Q, whole sub-branches 3-M and 2-B),
Blotchy mottle 1 (14) 5-A þ þ 2.8  107
several situations could be observed. (1) The c-PCR-positive,
Blotchy mottle 2 (66) 2-B þ þ 9.7  106
Blotchy mottle 3 (12) 5-A þ þ 6.4  106 blotchy mottle leaf samples, i.e. those with the highest liberibacter
Blotchy mottle 4 (110) 2-C þ þ 5.1  106 titers (red triangles), were frequently located at the distal ends of
Blotchy mottle 5 (8) 5-A þ þ 1.5  106 the lateral shoots. This was the case for all 4 samples of sub-branch
Blotchy mottle 6 (167) 2-D þ þ 1.1  106 2-B (Fig. 8B), 18 of 26 samples of sub-branch 3-M (Fig. 8C), 22 of 28
Zn deficiency 7 (172) 2-D þ þ 1.1  107
Zn deficiency 8 (6) 5-A þ þ 1.6  106
samples of sub-branch 4-Q (Fig. 8D), and all 9 samples of sub-
No symptoms 9 (195) 2-D þ þ 1.2  106 branch 5-A (Fig. 8E). In these cases, the ‘‘red circle’’ samples were
No symptoms 10 (10) 5-A þ þ 6.1  105 frequently found close to the distal ‘‘red triangle’’ samples, reflect-
Zn deficiency 11 (323) 3-M  þ þ 1.5  106 ing, locally, a relatively even distribution of liberibacters. (2) In
Zn deficiency 12 (468) 4-Q  þ þ 2.5  105
the case of shoots without blotchy mottle ‘‘red triangle’’ samples,
Zn deficiency 13 (470) 4-Q  þ þ 4.4  104
No symptoms 14 (363) 3-N  þ þ 4.1  106 the ‘‘red circle’’ samples replaced the ‘‘red triangle’’ samples at the
No symptoms 15 (361) 3-N  þ þ 1.1  105 distal end of the shoots. (3) It was also observed that samples neg-
No symptoms 16 (538) 4-Q  þ þ 9.5  104 ative by two (white circles) or three (green rings) PCR techniques,
No symptoms 17 (548) 4-Q  þ þ 8.0  104 i.e. with no or very few liberibacters, were located near blotchy mot-
No symptoms 18 (467) 4-Q  þ þ 7.9  104
No symptoms 19 (348) 3-M  þ þ 7.8  104
tle samples (red triangles), i.e. samples with high amounts of liber-
No symptoms 20 (372) 3-N  þ þ 7.1  104 ibacters (Fig. 8E, compare samples 11 and 12, 13 and 14, 22 and 23,
No symptoms 21 (581) 4-Q  þ þ 3.1  104 and 26 and 27), reflecting, locally, an uneven distribution of liberi-
No symptoms 22 (496) 4-Q  þ þ 2.7  104 bacters. Finally, symptomatic fruits could be found next to all possi-
No symptoms 23 (88) 2-B  þ þ 2.4  104
ble types of leaf samples, but most frequently close to ‘‘red triangle’’,
No symptoms 24 (294) 3-M  þ þ 1.7  104
No symptoms 25 (91) 2-B  þ þ 1.6  104 ‘‘red circle’’, and ‘‘blue circle’’ samples.
No symptoms 26 (320) 3-M  þ þ 1.5  104
No symptoms 27 (483) 4-Q  þ þ 1.5  104 4. Discussion and conclusions
No symptoms 28 (401) 3-N  þ þ 1.1  104
No symptoms 29 (316) 3-M  þ þ 8.6  103
No symptoms 30 (241) 3-J  þ þ 7.6  103
The sensitivity of the SYBR Green RTi-PCR technique for detec-
No symptoms 31 (402) 3-N  þ þ 6.5  103 tion of Ca. L. americanus in field samples is such that a positive
No symptoms 32 (331) 3-M  þ þ 4.9  103 amplification is still obtained with only 10 liberibacters per PCR
No symptoms 33 (81) 2-B  þ þ 4.2  103 reaction mixture. This sensitivity is at least as high as that of similar
No symptoms 34 (383) 3-N  þ þ 4.2  103
techniques published previously for the detection of the African,
No symptoms 35 (326) 3-M  þ þ 3.8  103
No symptoms 36 (374) 3-N  þ þ 3.7  103 Asian and/or American liberibacter [42–44]. Interestingly, nested
No symptoms 37 (523) 4-Q  þ þ 2.6  103 PCR with universal primer pair FD1/RP1 followed by primer pair
No symptoms 38 (378) 3-N  þ þ 2.2  103 GB1/GB3 (n-PCR), was almost as sensitive as RTi-PCR. In the case
No symptoms 39 (56) 2-B   þ 7.0  102 of the apple proliferation phytoplasma, nested PCR and RTi-PCR
No symptoms 40 (346) 3-M   þ 6.6  102
No symptoms 41 (321) 2-G   þ 4.6  102
had also similar sensitivities [45]. In comparison, conventional
No symptoms 42 (477) 4-Q   þ 2.6  102 16S rDNA PCR with primers GB1/GB3 (c-PCR) was 1000 times less
No symptoms 43 (335) 3-M   þ 2.3  102 sensitive. Yet, in routine testing for liberibacter in field samples, it
All 43 samples were used for quantitative RTi-PCR after samples 1–10 had been gave positive reactions with practically all leaf samples showing
found positive by c-PCR, samples 11–38, by n-PCR, and samples 39–43, by RTi-PCR. blotchy mottle, and has been most useful to confirm HLB in
thousands of citrus trees in SSP. Blotchy mottle leaves test positive
by c-PCR because they have the highest titers of liberibacters, and
by white circles, and those negative by all three PCR techniques, by this is true for all three liberibacter species. Quantitative RTi-PCR
green rings. Normal fruits are represented by yellow ellipses, and has shown blotchy mottle leaves to contain an average value of
symptomatic HLB fruits, by green ones. 107 liberibacters per gram of midribs (l/g). Even the lowest liberi-
Sub-branches 1-U and 1-V (Fig. 8A) are representative of bacter titers in blotchy mottle leaves are 10 times higher than the
symptomless branch 1, the leaves of which tested negative by either limit of detection of conventional PCR. This explains why leaves
two PCR techniques (white circles) or three PCRs (green rings). Sub- with blotchy mottle symptoms are the most reliable tissues for
branch 5-A (Fig. 8E) was the least affected one of the four symptom- HLB diagnostic. It seems as if high titers of liberibacters (106–
atic sub-branches of Fig. 8. The proximal end of the sub-branch 107 l/g) are required for citrus leaves to express blotchy mottle
(end closest to tree trunk), including the major lateral shoot, had symptoms. The highest liberibacter titer, 3.4  108 l/g, was seen in
only PCR-negative leaf samples, and the fruits on the major shoot mottled periwinkle leaves. This result confirms earlier EM observa-
were normal. However, its distal end had several samples positive tions showing that the number of HLB bacteria in the sieve tubes of
by c-PCR (red triangles and circles), including blotchy mottle leaves periwinkle leaves was higher than that in sweet orange sieve tubes
(red triangles). Sub-branch 4-Q (Fig. 8D), with 129 samples, was [15], and explains why liberibacter-infected periwinkle plants have
much larger than sub-branch 5-A, with 37 samples only, but it been useful in the development of cellular and molecular detection
resembled sub-branch 5-A in that it had many c-PCR-positive techniques (13, 14, 24, 30, 32), especially in the 1980s and early
samples (red triangles and circles) at the distal end, while at the 1990s, at times when PCR was not yet available.
proximal end, samples were essentially PCR-negative (white circles Among the 822 leaf samples collected on the symptomatic
and green rings) or n-PCR-positive (blue circles). Most symptomatic Westin sweet orange tree, all 111 leaf samples showing blotchy
fruits were located close to samples positive by c-PCR. Contrary to mottle symptoms were unambiguously positive by c-PCR. In
sub-branches 5-A and 4-Q, sub-branch 3-M (Fig. 8C) had many addition to leaves with blotchy mottle symptoms, as much as 84%
c-PCR-positive samples (red triangles and circles) at both proximal of leaf samples with zinc deficiency symptoms were PCR-positive.
 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150
Such leaves are the second best choice for HLB diagnosis, especially
when leaves with blotchy mottle are absent, as is the case in the
summer months in SSP as well as Florida. The association of zinc
deficiency symptoms with HLB has been pointed out long ago
[46]. Differentiation, in the orchard, between genuine, soil-induced
zinc deficiency and HLB-associated zinc deficiency is based on the
fact that the first will affect practically all trees in the orchard,
and each tree will have symptoms relatively well-spread all over
the canopy, while the second will affect only some of the trees in
the orchard, with trees showing yellow shoots only here and there
on the canopy.
Even symptomless leaves from the Westin sweet orange tree
gave positive PCR reactions. Of the 634 symptomless leaf samples,
81 samples gave positive amplifications with c-PCR, and 30
samples, while negative by c-PCR, were positive with n-PCR. The
latter samples had relatively low liberibacter titers, averaging
2.6  104, even though a few samples had higher liberibacter titers.
Finally, some symptomless leaf samples, negative by both c-PCR
and n-PCR, gave a positive amplification with RTi-PCR, and their
liberibacter titers were even lower, averaging 4.6  102 l/g. Accord-
ing to the sensitivity studies of Fig. 5, these titers should have been
detected not only by RTi-PCR, but also by n-PCR. This observation
suggests that RTi-PCR turns out to be somewhat more sensitive
than n-PCR to detect liberibacters in field samples. Similarly,
some samples with symptomless leaves or zinc deficiency leaves,
negative by c-PCR, but positive by n-PCR, had liberibacter titers
high enough to be detected by c-PCR. In these samples, the
presence of PCR inhibitors in the undiluted DNA preparation
generally used for c-PCR is very probably the reason for lower
sensitivity of c-PCR.
All PCR-positive samples came from the four symptomatic
branches, while all samples collected on the symptomless
branch were negative (Table 3). This result suggests that the
PCR-negative, symptomless branch 1 was either free of liberi-
bacters and not yet infected, or had liberibacter titers too low
to be detected even by RTi-PCR, in spite of the fact that the
four other branches of the tree were symptomatic and had
high titers of liberibacters in many of their leaves. This observa-
tion illustrates the uneven distribution of the liberibacters
within the tree.
On trees affected recently, leaf symptoms frequently start at the
distal end of a branch, on the outside of the canopy, probably near
the site where liberibacters have been inoculated by the psyllid
vectors. On two (Fig. 8D and E) of the four symptomatic branches
of the Westin sweet orange tree, leaves at the distal end of the
branches had much higher liberibacter titers than those at the
proximal end, near the trunk. This suggests that multiplication of
the liberibacters is more active at the distal end, where new shoots
are flushing periodically, than at the proximal end, inside the tree.
The lower liberibacter titers at the proximal ends could be due to
a lesser multiplication of the liberibacters or else to a slow invasion
of the proximal end by the HLB agents from the distal end.
RTi-PCR has recently been shown capable of detecting the Asian
liberibacter in symptomless trees of an affected orchard, the leaf
sample consisting of four fully expanded leaves collected arbitrarily
from three of the four sides of the tree [47]. However, many other
symptomless trees tested negative, some of which might have
been infected, but could have tested negative because of the
difficulty of choosing the right leaves, knowing that, even on
a symptomatic tree, some branches might still be infection-free
or have a very low liberibacter content. Obviously, the outcome of
the RTi-PCR tests depends not only on the sensitivity of the PCR
technique, but also on the type of leaves that are selected. For the
detection of infected but symptomless trees, using the sensitive
RTi-PCR technique is definitely an improvement over conventional
PCR, but leaf sampling remains a problem.
149
Acknowledgements
We would like to acknowledge the very capable assistance of
Patrick Bonnet (INRA, Bordeaux, France) in growing, and taking
care of, citrus seedlings and periwinkle plants in the greenhouse,
and of Ana Paula Leite Russo, Anelise G. Mariano, and Mateus A.
Santos (Fundecitrus, Araraquara, Brazil) in the many aspects of PCR.
References
[1] Bové JM. Huanglongbing: a destructive, newly-emerging, century-old disease
of citrus. J Plant Pathology 2006;88:7–37.
[2] Lin Kung Hsiang. Yellow shoot of citrus. Symptomatology. Investigations in the
cause of huanglongbing. Natural transmission and spread. General conclu-
sions. Acta Phytopathologica Sinica 1956;2:1–42.
[3] McClean APD, Oberholzer PCJ. Greening disease of sweet orange: evidence
that it is caused by a transmissible virus. South Afr J Agric Sci 1965;8:253–76.
[4] McClean APD, Oberholzer PCJ. Citrus psylla, a vector of the greening disease of
sweet orange. South Afr J Agric Sci 1965;8:297–8.
[5] Capoor SP, Rao DG, Viswanath SM. Diaphorina citri Kuway., a vector of the
greening disease of citrus in India. Indian J Agric Sci 1967;37:572–6.
[6] Martinez AL, Wallace JM. Citrus leaf mottle-yellows disease in the Philippines
and transmission of the causal virus by a psyllid, Diaphorina citri. Plant Dis Rep
1967;51:692–5.
[7] Etienne J, Aubert B. Biological control of psyllid vectors of greening disease on
Reunion Island. In: Calavan EC, Garnsey SM, Timmer LW, editors. Proceedings
of 8thconference of international organization of citrus virologists. Riverside,
CA: Department of Plant Pathology, University of California; 1980. p. 118–21.
[8] Aubert B, Quilici S. Biological control of the African and Asian citrus psyllids
(Homoptera: Psyllloidea), through Eulophid and Encyrtid parasites (Hyme-
noptera: Chalcidoidea) in Reunion Island. In: Garnsey SM, Timmer LW,
Dodds JA, editors. Proceedings of 9th conference of international organization
of citrus virologists. Riverside, CA: Department of Plant Pathology, University
of California; 1984. p. 100–8.
[9] McClean APD, Schwarz RE. Greening or blotchy-mottle disease of citrus.
Phytophylactica 1970;2:177–94.
[10] Garnier M, Bové JM. Citrus greening disease and the greening bacterium. In:
Moreno P, da Graça JV, Timmer LW, editors. Proceedings of 12th conference
of international organization of citrus virologists. Riverside, CA: Department
of Plant Pathology, University of California; 1993. p. 212–9.
[11] Varma A, Ahlawat YS, Chakraborty NK, Garnier M, Bové JM. Detection of the
greening BLO by electron microscopy, DNA hybridization and ELISA in citrus
leaves with and without mottle from various regions in India. In: Moreno P,
da Graça JV, Timmer LW, editors. Proceedings of 12th conference of interna-
tional organization of citrus virologists. Riverside, CA: Department of Plant
Pathology, University of California; 1993. p. 280–6.
[12] Laflèche D, Bové JM. Structures de type mycoplasme dans les feuilles d’or-
angers atteints de la maladie du greening. CR Acad Sci Paris 1970;270:1915–7.
[13] Garnier M, Danel N, Bové JM. Aetiology of citrus greening disease. Ann
Microbiol (Inst Pasteur) 1984:169–79.
[14] Garnier M, Danel N, Bové JM. The greening organism is a Gram negative
bacterium. In: Garnsey SM, Timmer LW, Dodds JA, editors. Proceedings of
9th conference of international organization of citrus virologists. Riverside,
CA: Department of Plant Pathology, University of California; 1984. p. 115–24.
[15] Garnier M, Bové JM. Transmission of the organism associated with citrus
greening disease from sweet orange to periwinkle by dodder. Phytopathology
1983;73:1358–63 (1 to 9).
[16] Jagoueix S, Bové JM, Garnier M. The phloem-limited bacterium of greening
disease of citrus is a member of the a subdivision of the Proteobacteria. Int J
Syst Bacteriol 1994;44:379–86.
[17] Jagoueix S, Bové JM, Garnier M. Comparison of the 16S/23S ribosomal
intergenic regions of ‘‘Candidatus Liberobacter asiaticum’’ and ‘‘Candidatus
Liberobacter africanum’’, the two species associated with Citrus huanglong-
bing(greening) disease. Int J Syst Bacteriol 1997;47:224–7.
[18] Garnier M, Jagoueix-Eveillard S, Cronje P, Le Roux H, Bové JM. Genomic
characterisation of a liberibacter present in an ornamental Rutaceous tree,
Calodendrum capense, in the Western Cape province of South Africa. Proposal
for a ‘‘Candidatus Liberibacter africanus subsp. capensis. Int J Syst Evol
Microbiol 2000;50:2119–25.
[19] Murray RGE, Schleifer KH. Taxonomic notes: a proposal for recording the
properties of putative taxa of prokaryotes. Int J Syst Bacteriol 1994;44:174–6.
[20] Garnier M, Bové JM, Cronje CPR, Sanders GM, Korsten L, Le Roux HF. Presence
of ‘‘Candidatus Liberibacter africanus’’ in the Western Cape province of South
Africa. In: da Graça JV, Lee RF, Yokomi RK, editors. Proceedings of 14th confer-
ence of international organization of citrus virologists. Riverside, CA: Depart-
ment of Plant Pathology, University of California; 2000. p. 369–72.
[21] Texeira DC, Ayres AJ, Kitajima EW, Tanaka FAO, Danet J, Jagoueix-Eveillard S,
et al. First report of a huanglongbing-like disease of Citrus in Sao Paulo State,
Brazil, and association of a new Liberibacter species, ‘‘Candidatus Liberibacter
americanus’’, with the disease. Plant Dis 2005;89:107.
[22] do Carmo Teixeira D, Dane JL, Eveillard S, Martins EC, de Jesus Jr WC,
Yamamoto PT, et al. Citrus huanglongbing in São Paulo State, Brazil: PCR
150 D.C. Teixeira et al. / Molecular and Cellular Probes 22 (2008) 139–150
detection of the ‘‘Candidatus’’ Liberibacter species associated with the disease.
Mol Cell Probes 2005;9:173–9.
[23] do Carmo Teixeira D, Saillard C, Eveillard S, Danet JL, Ayres AJ, Bové JM.
Candidatus Liberibacter americanus’’, associated with citrus huanglongbing
(greening disease) in São Paulo State, Brazil. Int J Syst Evol Microbiol 2005;
55:1857–62.
[24] Villechanoux S, Garnier M, Laigret F, Renaudin J, Bové JM. The genome of the
non-cultured, bacterial-like organism associated with citrus greening disease
contains the nusG-rplKAJL-rpoBC gene cluster and the gene for a bacterio-
phage type DNA polymerase. Curr Microbiol 1993;26:161–6.
[25] Planet P, Jagoueix S, Bové JM, Garnier M. Detection and characterization of the
African citrus greening liberobacter by amplification, cloning and sequencing
of the rplKAJL-rpoBC operon. Curr Microbiol 1995;30:137–41.
[26] Hocquellet A, Bové JM, Garnier M. Isolation of DNA from the uncultured
‘‘Candidatus Liberobacter’’ species associated with citrus huanglongbing by
RAPD. Curr Microbiol 1999;38:176–82.
[27] Bastianel C, Garnier-Semancik M, Renaudin J, Bové JM, Eveillard S. Diversity of
‘‘Candidatus Liberibacter asiaticus’’ based on the omp gene sequence. Appl
Environ Microbiol 2005;71:6473–8.
[28] Garnier M, Gao SJ, He Y, Villechanoux S, Gandar J, Bové JM. Study of the
greening organism (GO) with monoclonal antibodies: serological identifica-
tion, morphology, serotypes and purification of the GO. In: Brlanski RH,
Lee RF, Timmer LW, editors. Proceedings of 11thconference of international
organization of citrus virologists. Riverside, CA: Department of Plant
Pathology, University of California; 1991. p. 428–35.
[29] Gao S, Garnier M, Bové JM. Production of monoclonal antibodies recognizing
most strains of the greening BLO by in vitro immunization with an antigenic
protein purified from the BLO. In: Moreno P, da Graça JV, Timmer LW, editors.
Proceedings of 12th conference of international organization of citrus
virologists. Riverside, CA: Department of Plant Pathology, University of Califor-
nia; 1993. p. 244–9.
[30] Villechanoux S, Garnier M, Bové JM. Purification of the bacterium-like
organism associated with greening disease of citrus by immunoaffinity
chromatography and monoclonal antibodies. Curr Microbiol 1990;21:175–80.
[31] Garnier M, Bové JM. Distribution of the huanglongbing (greening) liberobacter
species in fifteen African and Asian countries. In: da Graça JV, Moreno P,
Yokomi RK, editors. Proceedings of 13th conference of international organiza-
tion of citrus virologists. Riverside, CA: Department of Plant Pathology,
University of California; 1996. p. 388–91.
[32] Villechanoux S, Garnier M, Renaudin J, Bové JM. Detection of several strains of
the bacterium-like organism of citrus greening disease by DNA probes. Curr
Microbiol 1992;24:89–95.
[33] Bové JM, Garnier M, Ahlawat YS, Chakraborty NK, Varma A. Detection of the
Asian strains of the greening BLO by DNA-DNA hybridization in Indian orchard
trees and Malaysian Diaphorina citri psyllids. In: Moreno P, da Graça JV,
Timmer LW, editors. Proceedings of 12th conference of international
organization of citrus virologists. Riverside, CA: Department of Plant Pathol-
ogy, University of California; 1993. p. 258–63.
[34] Korsten L, Jagoueix S, Bové JM, Garnier M. Huanglongbing (greening)
detection in South Africa. In: da Graça JV, Moreno P, Yokomi RK, editors.
Proceedings of 13th conference of international organization of citrus
virologists. Riverside, CA: Department of Plant Pathology, University of Califor-
nia; 1996. p. 395–8.
[35] Jagoueix S, Bové JM, Garnier M. PCR detection of the two ‘Candidatus’ libero-
bacter species associated with greening disease of citrus. Mol Cell Probes
1996;10:43–50.
[36] Hocquellet A, Toorawa P, Bové JM, Garnier M. Detection and identification of
the two ‘‘Candidatus Liberobacter sp.’’ associated with citrus huanglongbing
by PCR amplification of ribosomal protein genes of the beta operon. Mol
Cell Probes 1999;13:373–9.
[37] Coletta-Filho HD, Targon MLPN, Takita MA, De Negri JD, Pompeu JJ,
Machado MA. First Report of the causal agent of Huanglongbing (‘‘Candidatus
Liberibacter asiaticus’’) in Brazil. Plant Dis 2004;88:1382.
[38] Murray MG, Thompson WF. Rapid isolation of high molecular weight plant
DNA. Nucleic Acids Res 1980;239:487–91.
[39] Weisburg WG, Barns SM, Pelletier DA, Lane DJ. 16S ribosomal DNA amplifica-
tion for phylogenetic study. J Bacteriol 1991;173:697–703.
[40] Rajal VB, McSwain BS, Thompson DE, Leutenegger CM, Kildare BJ, Wuertz S.
Validation of hollow fiber ultrafiltration and real-time PCR using bacterio-
phage PP7 as surrogate for the quantification of viruses from water samples.
Water Res 2007;41(7):1411–22.
[41] Seurinck S, Defoirdt T, Verstraate W, Sciliano SD. Detection and quantification
of the human-specific HF183 Bacteroides 16S rRNA genetic marker with real-
time PCR for assessment of human faecal pollution in freshwater. Environ
Microbiol 2005;7:249–59.
[42] Li W, Hartung JS, Levy L. Quantitative real time PCR for detection and
identification of Candidatus Liberibacter species associated with citrus huan-
glongbing. J Microbiol Methods 2006;66:104–15.
[43] Li W, Hartung JS, Levy L. Evaluation of DNA amplification methods for
improved detection of ‘‘Candidatus Liberibacter species’’ associated with citrus
huanglongbing. Plant Dis 2007;91:51–8.
[44] Wang Z, Yin Y, Hu H, Yuan Q, Peng G, Xia Y. Development and application of
molecular-based diagnosis for ‘Candidatus Liberibacter asiaticus’, the causal
pathogen of citrus huanglongbing. Plant Pathol 2006;55:630–8.
[45] Torres E, Bertolini E, Cambra M, Monton C, Martin MP. Real Time PCR for
simultaneous and quantitative detection of quarantine phytoplasmas from
apple proliferation (16SrX) group. Mol Cell Probes 2005;19:334–40.
[46] Lee HA. The relation of stocks to mottled leaf of citrus leaves. Philipp J Sci
1921;18:85–95.
[47] Irey MS, Gast T, Gottwald TR. Comparison of visual assessment and polymerase
chain reaction assay testing to estimate the incidence of the huanglongbing
pathogen in commercial Florida citrus. Proc Fla State Hort Soc 2006;119:89–93.